1. Field of the Invention
The present invention relates to a method and device for measuring welded bead cutting shapes in electric resistance welded pipes, including the position of the bead.
2. Description of the Related Art
In general, electric resistance welded pipes, such as electric resistance steel pipes, are manufactured by continuously forming steel plates or steel coils into a tubular form, and continuously performing butt welding both sides of the steel coil in the longitudinal direction by means such as high-frequency induction heat-pressure welding or resistance heat-pressure welding.
A raised portion called a “bead” is formed at the weld portion of the electric resistance welded pipe due to the butt welding, on the inside of the pipe and the outside of the pipe. Generally, the bead is continuously cut in the longitudinal direction of the steel pipe by a cutting tool on the production line further downstream from the welder. The shape of the surface of the pipe after the bead portion has been cut off (hereafter referred to as “bead cutting shape”) preferably becomes one with the outline shape of the pipe other than the weld, such that where the bead was cannot be distinguished, ideally. In order to achieve this, the cutting tool tip must be held at an appropriate position on the surface of the electric resistance welded pipe.
Conventionally, a worker has measured the bead cutting shape, visually or using a micrometer or the like, at the time of starting cutting, and adjusted the cutting tool to the optimal position. However, there have been cases wherein, while manufacturing a number of electric resistance welded pipes, the position of the cutting tool shifts due to a variety of reasons, the blade of the cutting tool is nicked, etc., causing defective cutting such as not all of the bead being cut off of the electric resistance welded pipe product or cutting too deep. Such cutting defects not only mar the look of the electric resistance welded pipe product; using electric resistance welded pipes with cutting defects for piping subjected to pressure such as gas lines or the like may place the pipe at risk of rupturing.
Accordingly, there is the need to measure and monitor the bead cutting shape while manufacturing, and suitably correct the cutting tool position or replace the cutting tool with a new one, according to the results thereof.
Even with regard to the outer face of the pipe which is readily observed from the outside, monitoring of the bead cutting shape has to depend on visual observation of workers, so precision and reproducibility thereof is insufficient, and there have been quantitative and reliability problems.
With regard to the inner face of the pipe, the bead cutting portion cannot be directly observed during manufacturing, due to the configuration of the production line. Accordingly, the end portion of the pipe is observed at the process where the pipe is cut at the end position of the line. Alternatively, the line is stopped to cut out a sample of the bead position of the pipe by gas cutting, and the inner face is observed. With the former method, the observation position is several tens of meters downstream from the cutting position, so there has been a problem in that in the event that an abnormality occurs in the cutting, the defective pipe becomes longer until detected, resulting in decrease in yield. Also, with the later method, the cutting tool seizes due to friction heat in the event that the line is stopped, so the cutting tool must be retracted. Resetting the cutting tool and restarting the line creates a step between the previously-cut bead cutting shape and the new cutting shape, so that portion is unusable as a product. There has been the problem that productivity decreases due to stopping the line for taking the sample. Only a part in the longitudinal direction of the product can be inspected in both of the above-described methods, so there has been the problem that the quality cannot be guaranteed for the entire length of the product.
In order to solve these problems, a bead cutting shape measurement method using the optical cutting method has been proposed. Examples of the optical cutting method are disclosed in Japanese Unexamined Patent Application Publication No. 57-108705 and Japanese Examined Patent Application Publication No. 60-1138. As shown in FIG. 13, slit light 121 from a light source 120 is cast on a measuring object (electric resistance welded pipe 110), and this is observed from a different angle with a camera 130, thereby observing a slit image (optical cutting image) deformed following the surface shape of the measuring object 110. The shape of the object can be calculated from this optical cutting image and the geometric position of the observation optical system. This is advantageous in that the observation optical system is simple, that the measurement sensitivity can be widely changed according to the geometric placement of the observation optical system, and so forth. The region outside of the irradiation region of the slit light is called “background texture”.
Japanese Unexamined Patent Application Publication No. 52-96049 proposes a bead shape observation method wherein an uncut weld bead portion is observed with the optical cutting method, and marks corresponding to an enlargement ratio determined by optical placement on a display monitor 140.
However, these methods only display a measurement image. Judgment of the bead cutting shape is performed by workers visually judging the monitor 140, and automatic measurement is not yet employed.
An example of a determination method for automatic measurement is the technique disclosed in Japanese Patent No. 2618303, for example. According to this, a method is proposed wherein a picture is taken of the steel pipe bead cutting portion using an optical cutting image from slit light and an ITV camera at the time of measuring the shape following cutting the welding bead of the electric resistance welded pipe, and as shown in FIG. 13, the cross-sectional shape is calculated by performing thinning processing (taking a region wherein one pixel is connected in one direction as a thin line) on the cross-sectional shape picture, the cut portion and uncut portion are distinguished by the luminance of the cross-sectional shape, the value for the center of the distinguished cut portion and the value at the right edge of the cut portion and the value at the left edge of the cut portion are obtained, and the cutting depth, and amount of cutting inclination are calculated based on these three calculated values.
However, with the technique disclosed in Japanese Patent No. 2618303, the specific method for thinning processing only involves performing computation for directly substituting luminance to Y-axial coordinates, such as plotting the maximum luminance in a direction parallel to the pipe axis (Y-axial direction) obtained from the optical cutting image onto the coordinates on an X axis extending in the circumferential direction of the pipe (this is equivalent to the width direction for material steel plates and steel coils, and accordingly with hereafter be referred to as the width direction), so there is the problem in that there are cases wherein an accurate cross-sectional shape cannot be obtained.
Describing this in detail, from the experience of the Inventors repeating experiments at the manufacturing site, while the surface of the cut portion of the electric resistance welded pipe immediately following cutting has a specular surface, the surrounding uncut portions are blackish due to an oxide layer adhering thereto, so the degree of diffusion of the slit light thereof differs. Accordingly, the luminance of the optical cutting image of the bead cutting portion is not necessarily around the same degree in the width direction. For example, there are cases such as in FIG. 14 wherein almost all of the slit light of the cut portion exhibits specular reflection (reflection in the opposite direction to the incident direction at the same angle as the incident angle), with the luminance thereof being one-tenth or less than that of the uncut portion. This is due to the fact that in the event that the incident angle and the reception angle differ, such specular reflection light actually appears to have less luminance.
In such cases, the optical cutting image becomes lost in noise, so the bead cutting shape cannot be obtained well. Attempting to raise the luminance of the cut portion by raising the gain or extending the exposure time of the observation optical system of the ITV camera or the like leads to exceeding the range of the maximum luminance of the specifications of the observation optical system such as the aforementioned camera or the like (halation) at the uncut portion as indicated by (c) in FIG. 1, so the shape of the uncut portion cannot be accurately distinguished. The reason is that in the event that such exceeding of the range of luminance occurs, multiple pipe axis direction coordinates (Y-axial coordinates) indicating the maximum luminance appear at the uncut portion on the optical cutting image, and the pipe axis direction coordinates (Y-axial coordinates) indicating the maximum luminance cannot be uniquely determined.
The present invention has been made to solve the problems such as described above, and accordingly, it is an object thereof to provide a method and device for precisely measuring the bead cutting shape of electric resistance welded pipe without being affected by the difference in luminance level between the cut portion and uncut portion in optical cutting images.
With regard to suppressing the effects of noise in such an optical cutting method, proposals in other fields using the optical cutting method are conventionally known.
Japanese Unexamined Patent Application Publication No. 57-208404 proposes a method wherein an optical cutting image is searched vertically from top to bottom and an optical cutting line is extracted only from within a section wherein a portion greater than a predetermined setting value first occurs, and subsequent extraction of optical cutting lines is terminated at the one scanning line, thereby preventing erroneously detecting abnormal reflections at portions of the optical cutting line on the object other than the slit luminescent line position.
Also, Japanese Unexamined Patent Application Publication No. 2-35306 proposes a shape detecting method wherein the entire region of the acquired optical cutting image is scanned in the direction crossing the optical cutting line, and in the event that there is a peak value due to the noise image on the scanning line, an optical cutting search range is set based on the optical cutting line position detected on the same scanning line over the entire screen, thereby ignoring noise.
Also, Japanese Unexamined Patent Application Publication No. 4-240508 proposes a three-dimensional shape recognition device for calculating coordinates of an object of measurement based on an optical cutting image, and judging the image to be an unreal image in the event that the shape thereof exists in a needle-like shape separated from surrounding images, thereby ignoring that data and recognizing the shape.
However, with the method disclosed in Japanese Unexamined Patent Application Publication No. 57-208404, a fixed threshold V1 is used for recognizing the only optical cutting extraction section, so application thereof to measuring bead cutting shapes wherein the luminance of the optical cutting line of the acquired optical cutting image changes greatly between the cut portion and uncut portion is impossible, as already described.
Also, the method disclosed in Japanese Unexamined Patent Application Publication No. 2-35306 assumes that parts having raised shapes with a generally uniform size in the longitudinal direction are arrayed in the direction of the slit light at equal intervals, as with soldered portions of electronic parts, and generation of noise is explained as being secondary light due to reflected slit light off of a neighboring part, so this is clearly inapplicable to the problems of the present invention since a similar noise generation state is not generated at the cut portion of an electric resistance welded pipe.
Also, with the technique disclosed in Japanese Unexamined Patent Application Publication No. 4-240508, in the state that the optical cutting image of the cut bead becomes non-continuous at a step portion, the step portion is erroneously recognized as a non-continuous image (unreal image) and ignored, so there has been the problem that this could lead to missing cutting defects.
That is to say, no method has yet been found in the Technical Field of the Present Invention or the field of shape recognition techniques by optical cutting in other Technical Fields, for accurately measuring weld bead cutting shapes even in the event that the SN ratio between the optical cutting line image and the surroundings deteriorates.
The present invention has been made to solve the above-described problems, and accordingly it is an object thereof to provide a measurement method and measurement device which can readily recognize image processing abnormality portions due to deterioration in the SN ratio in three-dimensional shape measurement by the optical cutting method as being uneven portions due to cutting.
With conventional inventions relating to steel pipe welding bead detection methods or devices, mechanical methods, methods using eddy current sensors, optical methods, and so forth, have been proposed.
(Patent Document 1)
Japanese Examined Patent Application Publication No. 59-2593
(Patent Document 2)
Japanese Unexamined Patent Application Publication No. 2000-176642
(Patent Document 3)
Japanese Unexamined Patent Application Publication No. 5-133940
(Patent Document 4)
Japanese Unexamined Patent Application Publication No. 5-18904
(Patent Document 5)
Japanese Unexamined Patent Application Publication No. 9-72851
(Patent Document 6)
Japanese Unexamined Patent Application Publication No. 60-135705
As for a mechanical method, Patent Document 1, for example, proposes a method for detecting deviation of the outer face weld portions of the running pipe using contact-type rollers.
Also, as for a method using a eddy current sensor, Patent Document 2 proposes a center position detection method for the welding bead wherein a detecting head comprising a magnetic core, performing uniform circular motion on a concentric axis within a transmission coil and reception coil disposed in a concentric cylindrical fashion, is erected above the welding bead and brought into close proximity, and at the time of the magnetic core passing over the welding bead, passage timing is detected twice per rotation of the magnetic core based on the change in impedance of the reception coil, with the time passing between these passage timings being computed and compared, thereby detecting the center position of the welding bead.
Also, Patent Document 3 proposes a method wherein a plurality of eddy current sensors are disposed in a form generally extending in one direction, and scanning the object of detection by sequentially switching the roles of the eddy current sensors between magnetization, induced-magnetization, and detection, thereby detecting the bead position from the detection waveform.
Also, as for optical methods, methods have been proposed such as the method disclosed in Patent Document 4 wherein an image of the surface of the pipe is taken and signal waveform features inherent to the weld portion and base metal are extracted and checked against features stored beforehand so as to distinguish these, or a steel pipe welding bead detection method such as disclosed in Patent Document 5 wherein a steel pipe is rotated in the circumferential direction and an image is taken of the surface of the pipe with an ITV camera or the like while irradiating sector light on the surface of the pipe or irradiating sector light consisting of a point light being scanned, the picture signals are subjected to noise removal and tilt correction and the like to form an image which is subjected to correction processing, following which circular arc application is used wherein the difference between a circular arc image to which a circular arc has been applied and an actual image is obtained based on the corrected image, and in the event that the difference data exceeds a preset threshold value, judgment is made that to be a welding bead, while also checking against a tolerance range for the welding bead width set beforehand for a width range exceeding the threshold value, thereby determining the bead position.
Also, Patent Document 6 proposes a bead shape automatic measuring device as a common welding bead position and shape automatic measurement technique, wherein an image is taken of the welding bead from above and from the side, analog image information from the image-taking unit is converted into grayscale level digital information, and the width and height and the like of the welding bead is detected based on the digital image information.
However, with the method using the contact-type sensor such as disclosed in Patent Document 1, the height of the bead must be approximately constant in the longitudinal direction with the irregularities in height thereof being relatively steep, so in the event that the irregularities in height of the bead is constantly smooth, in the event that the bead height is low, or in the event that the bead height is not constant in the longitudinal direction, accurate detection cannot be made.
Also, with the method using the concentric motion magnetic core within the concentric cylindrical coils disclosed in Patent Document 2, in the event that a twisted portion in the seam passes through the detection device position while transporting the electric resistance welded pipe which is the subject, or in the event that meandering occurs therein, the positional relation between the electric resistance welded pipe which is the subject and the concentric cylindrical coils or the concentric motion magnetic core which make up the detection device is offset, so accurate welding bead position detection cannot be performed.
Also, the method disclosed in Patent Document 3, wherein magnetization, induced-magnetization, and detection coils are disposed generally in one direction, readily responds to foreign material adhering to the surface of the pipe and irregularities and so forth in the height of the surface of the pipe besides the bead, so it is difficult to avoid erroneous detection, and also, in the event of dealing with various sizes such as with electric resistance welded pipes, multiple detecting heads must be prepared to handle the difference in shape thereof, which increases manufacturing costs.
Also, as for a common problem with these eddy current methods, there is the problem that separate shape measurement means must be provided for evaluating the shape of the welding bead regarding which position detection has been performed, increasing the manufacturing costs of the device.
In comparison with the above-described methods, the optical methods such as disclosed in Patent Document 4 and Patent Document 5 allow non-contact detection, and are advantageous in that not only bead position detection but also bead shape evaluation can be made with the same device configuration. However, there have been various problems in the above-described conventional optical methods.
That is, with the method disclosed in Patent Document 4 for weld-portion/base-metal features extraction, primary methods involve detecting the difference in brightness of the bead portion and other portions (base metal: hereafter referred to as “base pipe”), but the brightness (luminance) of the bead portion greatly depends on the welding conditions and the thickness of the base pipe, so beside difficulty in obtaining stable detection, there has been the problem of cases wherein bead recognition cannot be performed in the event that the luminance of the bead portion is low, in particular.
Also, with the steel pipe rotation circular arc application method disclosed in Patent Document 5, the steel pipe must be rotated in the circumferential direction, but at the welding stage of the electric resistance welded pipe, the steel pipe is often connected to the steel coil which is the base metal, making rotation impossible, and in addition to this problem, two circular arcs are calculated from four points of data at the image processing stage, so even in the event that noise processing is carried out, this method is readily affected by jagged shapes often observed in image data, so there is the problem that error readily occurs in the bead position that is calculated, and further, the roundness of the electric resistance welded pipe which is the subject is seldom poor, so there has been the problem that there is a limit to suppressing occurrence of detection error with this method which uses the geometrical principle of a circle wherein the center of the pipe exists on a perpendicular bisector of two points.
Also, with the camera image-taking method in Patent Document 6, a point where the gradiation for one line of image rapidly changes is searched as the method for determining the bead position, so there has been the problem that in the event that the luminance of the bead portion is low, or depending on the surface properties, there are cases wherein the bead position cannot be determined.
Further, besides these detection methods, as described in Patent Document 5, one skilled in the art would readily conceive a method for measuring the profile of the surface of the steel pipe including the bead position, using an optical cutting method or an optical distance measurement method, and detecting the bead position by processing the profile data. However, in this case, some sort of derivation processing assuming that sudden changes occur in the profile at the bead portion would be commonly applied as a method for processing the profile data, but advances in welding technology in recent years have led to smooth slopes on the bead, while such derivation processing accentuates minute noise which readily occurs in optical profile measurements, so detection of the bead position actually becomes more difficult.
The present invention has been made to solve the problems of the conventional art described above, and it is an object thereof to provide a method and device for accurately detecting the bead shape from the shape data of electric resistance welded pipes with the so-called optical cutting method detected by slit light or point light scanning, without effects of luminance or profile data noise.
Also, as for an optical method, as disclosed in Patent Document 7, a metal flow angle measurement method for electric resistance welded pipes has been proposed, wherein slit light is irradiated before bead cutting on a base pipe which is moving and the optical cutting profile obtained thereby is optically received as image, the width and height of the bead at the welded portion is detected from the obtained optical cutting profile reception signals, and the metal flow angle of the welded portion is computed based on the detection values of the width and height of the bead thus obtained.
(Patent Document 7)
Japanese Examined Patent Application Publication No. 60-7586
Also, Patent Document 8 proposes a metal flow angle measurement method for the welded portion of electric resistance welded pipes, wherein slit light is irradiated before bead cutting on a base pipe which is moving and optical cutting profiles each obtained thereby are optically received as an image, the surface position of the bead corresponding to a predetermined height within the range of ¾ to ⅓ of the maximum height of the bead based on the rising position of the beard at the welded portion is detected from the obtained optical cutting profile image reception signals, and the metal flow angle of the welded portion is computed based on the horizontal distance from the slope of the bead on the surface and the predetermined height thereof, corresponding to the predetermined height thus obtained.
(Patent Document 8)
Japanese Examined Patent Application Publication No. 60-25234
However, in the event of using a contact-type roller such as disclosed in Patent Document 1 and a speedometer together, the height of the bead must be approximately constant in the longitudinal direction with the irregularities in height thereof being relatively steep, so in the event that the irregularities in height of the bead is constantly smooth, in the event that the bead height is low, or in the event that the bead height is not constant in the longitudinal direction, there has been the problem that accurate detection cannot be made.
Also, with Patent Document 7, the shape of the welding bead is taken as being a trapezoid in form, with the relation between the width and height ratio thereof and the metal flow angle being calculated by a shape index computation circuit based on experiment expressions, but advances in welding technology in recent years have led to smooth slopes on the bead, and the optimal slope angle changes depending on the thickness of the plate or the usage, so there has been the problem that operating while experimentally switching over the calibration curve for each case becomes extremely troublesome.
Also, with the Patent Document 8, bead width information is used at ¾ and ⅓ of the bead height, so in addition to the aforementioned problems, there has been the problem that in the event that the bead shape is off of a triangular shape or trapezoid shape, e.g., in the event that the portion of ⅓ through ¾ of the height is vertical, the denominator of the metal flow computation is zero, the computation results are abnormal.
Also, a method may be conceived wherein the cross-sectional-direction shape (in the direction perpendicular to the axis) of the pipe including the weld bead portion is detected and the bead position and slope angle is calculated by derivative values thereof, but in the event that there is noise on the detected shape data, this is accentuated by the derivation computations of such a method, leading to the problems of erroneous bead shape detection or increased error in the slope angle calculations.
The present invention has been made to solve the problems in the conventional art as described above, and it is an object thereof to detect bead shapes with precision from the shape data of electric resistance welded pipes detected by the optical cutting method.